460ex and dithering

[nightster]

On the 460ex or any low noise camera, How important is dithering?

[NickK]

It depends on what you use it for - noise reduction, or, detail recovering.

Dithering - i.e. moving the mount by a certain amount to allow the captured images to then be aligned and show any noise by corresponding movement between frames - is also useful for recovering detail from under sampled images through drizzling of those original low noise frames. It's preferable to dither the shots in both DEC and RA in this instance.

There is also another useful reason for dither - recovering more accurate colour information for OSC cameras with a drizzle designed for colour rather than resolution.

[nightster]

So if your sampling is correct for seeing conditions and equip, and your mono, is it unnecessary?

[NickK]

So if your sampling is correct for seeing conditions and equip, and your mono, is it unnecessary?

I you want to stick with the norm for seeing conditions then you could. It's a subjective thing in the end - my preference in for smaller pixels to give over sampling and then let processing attempt to recover more detail from it. Some may see that as foolhardy, but that's my preference I'll run 0.88 arcsec/pixel where the seeing may be 1-2 arcsecs with a scope of 1.10 arcsec resolution - the only side effect is softening.

[Steve 1962]

This is quite a good read on the subject...

http://www.cyanogen.com/help/maximdl/Dark_Frame_Calibration.htm

I dither and dark frame calibrate my mono 460ex because in my simple mind, any dark frame subtraction will replace the hot pixels with a value that is unlikely to be the "true" value of the what the pixel would have recorded had it not been hot.

By dithering and using a sigma clip in the stacking routine - if the replaced value is a long way from the "true" value it'll be pulled back into line.

I suppose that, using this logic, it's valid to wonder whether dark frames are necessary if you dither and sigma clip!

[nightster]

What is a defect map and how does that fit in to the process?

[kalasinman]

I you want to stick with the norm for seeing conditions then you could. It's a subjective thing in the end - my preference in for smaller pixels to give over sampling and then let processing attempt to recover more detail from it. Some may see that as foolhardy, but that's my preference I'll run 0.88 arcsec/pixel where the seeing may be 1-2 arcsecs with a scope of 1.10 arcsec resolution - the only side effect is softening.

Sorry if a stupid question, learning..I can see how oversampling the image can soften it, but I fail to see how guiding oversampling could anything but improve guiding accuracy. Can you help me with this?

[NickK]

Sorry if a stupid question, learning..I can see how oversampling the image can soften it, but I fail to see how guiding oversampling could anything but improve guiding accuracy. Can you help me with this?

If you're using a centroid mass, you can either use the interpolated mass or the real mass. For the OSC, over sampling would mean a higher accuracy as the colours are closer together.